Table 2.
Balance of patient characteristics before and after IPCW, using a 15% threshold.
| Characteristic | Absolute standardized mean difference, % a | |||||
|---|---|---|---|---|---|---|
| Overall population | NOVA study-like subpopulation | |||||
| Before IPCW | After IPCW | After stabilized IPCW | Before IPCW | After IPCW | After stabilized IPCW | |
| Age group at index | 26.4 | 6.4 | 6.9 | 30.0 | 2.0 | 4.2 |
| Race | 28.4 | 11.2 | 11.0 | 25.8 | 13.8 | 14.0 |
| Practice type b | 30.6 | 15.8 | 15.3 | 27.8 | 14.7 | 14.7 |
| Region | 24.2 | 15.1 | 14.2 | 47.0 | 10.7 | 12.6 |
| ECOG performance status | 31.7 | 12.9 | 12.2 | – | – | – |
| Disease stage at diagnosis | 6.8 | 6.3 | 5.5 | 27.7 | 2.2 | 4.7 |
| Epithelial histology | 5.3 | 2.8 | 2.2 | 5.2 | 1.3 | 0.5 |
| Duration between the end of 1L and the start of 2L | 73.1 | 5.1 | 10.6 | 3.2 | 4.8 | 9.3 |
a
Standardized mean difference is calculated as the absolute difference in means, mean ranks, or proportions divided by the pooled standard deviation. For categorical variables, the overall standardized difference was calculated using Mahalanobis distance. 17
b
Practice type had a standardized mean difference of 15.3% in the overall population after stabilized IPCW, but the balance was achieved at the categorical level.
1L, first-line; 2L, second-line; ECOG, Eastern Cooperative Oncology Group; IPCW, inverse-probability-of-censoring weights.